PS3 (Research/Experimental) - NEC/TOKIN Capacitors Replacement - YLOD

[RIP-Felix]

Hi guys,
for the first time I got a YLOD on my ps3 80gb cechk-04 (supposedly with 65nm rsx). It's a weird YLOD though. It happened to me the first time while playing and the ps3 suddenly shut down. After leaving it off for a few days I tried to turn it back on. At the first attempts it turns off within a few seconds (with a yellow light and then a flashing red light) while after a few attempts the ps3 manages to turn on. Every time I turn it on the ps3 stays on a little longer until I can access the XMB. I even managed to jailbreak it in these conditions, if it is left on for some time it becomes "almost" stable. But when I open a game or do something heavy, the ps3 turns off within a few seconds (at most 1 minute). I installed "PS3 Advanced Toolset" and downloaded the SYSCON logs. The only error that comes up is "a0801002". What do you think is the problem? The RSX or the Nec/Tokins?
Thanks!

RSX tokins. That model, those symptoms, and specific error all point to this diagnosis.

Would you please post the usage time for postarity? You can get it from bgtoolset system management at the bottom under errors. Same page with the Flash dump, just at the bottom. Just don't post an image of the IDPS or PSID as they could be stolen and get you banned from PSN. So if you take a pic of that screen blur those out.

 

[imBenjamin741]

RSX tokins. That model, those symptoms, and specific error all point to this diagnosis.

Would you please post the usage time for postarity? You can get it from bgtoolset system management at the bottom under errors. Same page with the Flash dump, just at the bottom. Just don't post an image of the IDPS or PSID as they could be stolen and get you banned from PSN. So if you take a pic of that screen blur those out.

After twenty minutes I finally managed to turn it on...
This is the usage time: Statistics: Boot Count 2295 - Shutdown Count 2194 - Runtime 116d 8h 11mn 24s
It must be said that I left it unused for 2 or 3 years before turning it on again and after 30 minutes of playing it suddenly turned off. Nothing had ever happened to it before. It was perfect! It's probably more the age than the use itself.
Anyway thanks a lot RIP-Felix. I watched all your videos and learned a lot. Would you recommend that I replace the nec/tokins even though I don't have much experience with soldering? I would use tantalizers and 2R5TPE470M7 or 2R5TPE470M9 capacitors, it depends on the price at which I can find them. Otherwise do you know any shop in Europe that does these jobs? Thanks again

 

[RIP-Felix]

Any low ESR 7343 cap will do. There are aluminum polymer caps with 2000hr rating that are 3mohms each, for example. The TaPol M7's are better for ESR into the higher frequency, but the lower overall ESR of the AlPol may out weigh that...I still need to do a head to head oscilloscope test on several combinations to see for sure. It's splutting hairs for performance gains nobody will notice. So long as you don't exceed 10mOhms per cap (and use 12 of them to divide the combined ESR) then it'll be accemptable.

So K models "should" have a DIA-002 with the 1st revision 65nm CXD2982 RSX. I have since confirned these do have fixed underfill (Higher Tg). They could still lack the polyamide stress layer that protects the dielectric layers of the silicon (the actual nanoscopic circuits on the die), but it apparantly doesn't matter even if they do. The 65nm RSX is a tank as far as I can tell. I just pulled 2x 2982's from a lot of DIA-002's which both had over 1000 days of use on them and the errorlog was the same as yours (full of 1002s). Frankly, I'm proul of those tokins for holding out for so long before failing! 1000x 24hr days is 24000 hours. Which is 2.4 times the 10,000 hours I predicted for them to. With a bit more data about the average time they last, I can solve the math to predict the average temp they see. And then I can test what adding a thermal pad to them would do and run the math again to predict how long that mod could extend their lifespan. Which is one if the reasons I asked for the useage time. Thanks for that.

I know several people in europe that might be able to help. I'll DM you.

 

[Thibobo]

It is an RSX issue most likely, but reballing can't fix bumpgate. Since we can replace the RSX with a 65nm or 40nm RSX that are not affected, that is now the reccomended repair.

Thanks!
I might buy a reflow station, and learn the Frankenstein method. Which station do you recommend?

I'd like to make the ultimate ps3 from this one, it would be delid 40nm rsx + cell delided & also reballed with leaded solder
All tokins replaced with tantalizers

Is there any guide i can use for the Frankenstein process?


Edit: i found your tutorial... Big ups for you i'm just now interested in things you already were doing some time ago!.

https://www.psx-place.com/threads/f...cecha-with-40nm-rsx.28069/page-76#post-324546

 

[RIP-Felix]

Thanks!
I might buy a reflow station, and learn the Frankenstein method. Which station do you recommend?

Depends on your price range and goals. If you just want 1 PS3 and do not intend to continue reworking electonics in the future, then it's cheaper to pay someone like @Computer Booter for the service.

If you want the easiest/cheapest, but still capable option, then get and AHCI SC PRO.

If your budget allows, are serious about rework and want a superior machine, then an LY G580 is going to be better.

What I did I wouldn't reccomend!

I'd like to make the ultimate ps3 from this one, it would be delid 40nm rsx + cell delided & also reballed with leaded solder
All tokins replaced with tantalizers

There is no need to delid the RSX. Just paranoia. Once the 90nm is replaced with one that doesn't have a defect, there is no more temperature related failure. It is in a HS that's over designed for it's TDP anyway.

The issue with delidding is needing to readhere the IHS back on and people are simply not doing it. It is unwise to perform a mod out of paranoia about an issue that doesn't exist, and then leave the console in a state that is likely worse.

The CPU is anoter story. IMO most 90nm CELLs need delidded. But when you do, you should readhere the IHS with new silicone.

 

[truemaster]

reapply silicon to cell ihs is sure important. without all the pressure will be on the die. but rsx has rams on the corner so glue it back will save nothing

 

[Thibobo]

Depends on your price range and goals. If you just want 1 PS3 and do not intend to continue reworking electonics in the future, then it's cheaper to pay someone like @Computer Booter for the service.

If you want the easiest/cheapest, but still capable option, then get and AHCI SC PRO.

If your budget allows, are serious about rework and want a superior machine, then an LY G580 is going to be better.

What I did I wouldn't reccomend!


There is no need to delid the RSX. Just paranoia. Once the 90nm is replaced with one that doesn't have a defect, there is no more temperature related failure. It is in a HS that's over designed for it's TDP anyway.

The issue with delidding is needing to readhere the IHS back on and people are simply not doing it. It is unwise to perform a mod out of paranoia about an issue that doesn't exist, and then leave the console in a state that is likely worse.

The CPU is anoter story. IMO most 90nm CELLs need delidded. But when you do, you should readhere the IHS with new silicone.

Thanks for you honest opinion!

Looks like i might go with the LY IR8500, it seems better than the IR6500, and overall the best price/quality option out there.

Because i'm super motivated to learn this, if i'm confident enough i might make more than 1 frankenstein and/or offer the service in Belgium.

Do you have any recommendations for stencils? Direct heat ones are out of the question for me but i can't find a proper source or bundle/set on aliexpress for the 90x90 ps3 ones…

If there exists a bundle for multiple consoles like ps3,4,5… or even xbox ones i'd rather buy that than only the few for ps3. They moght also come in handy.

I also have a question about bottom support for the ps3 board, do you recommend bottom support or isn't it needed? If so is the standard bottom support enough or do i need a ps3 board bracket to secure the motherboard?

thnx for the feedback!

 

[haxxxen]

The issue with delidding is needing to readhere the IHS back on and people are simply not doing it. It is unwise to perform a mod out of paranoia about an issue that doesn't exist, and then leave the console in a state that is likely worse.

haven't done this, but using PC heatsinks with LM, which also has some adhesive effect. though, my "BastelStation" has its heatsinks upwards now and I have insulated both processors.

 

[RIP-Felix]

Thanks for you honest opinion!

Looks like i might go with the LY IR8500, it seems better than the IR6500, and overall the best price/quality option out there.

Because i'm super motivated to learn this, if i'm confident enough i might make more than 1 frankenstein and/or offer the service in Belgium.

Do you have any recommendations for stencils? Direct heat ones are out of the question for me but i can't find a proper source or bundle/set on aliexpress for the 90x90 ps3 ones…

If there exists a bundle for multiple consoles like ps3,4,5… or even xbox ones i'd rather buy that than only the few for ps3. They moght also come in handy.

I also have a question about bottom support for the ps3 board, do you recommend bottom support or isn't it needed? If so is the standard bottom support enough or do i need a ps3 board bracket to secure the motherboard?

thnx for the feedback!

Pay attention to the preheater working area. The AHCI SC Pro has a much bigger area than similar looking ones.

 

[RIP-Felix]

reapply silicon to cell ihs is sure important. without all the pressure will be on the die. but rsx has rams on the corner so glue it back will save nothing

The die is 0.2mm taller than the VRAM modules. The thermal plaster fills this gap. Without the plaster all the pressure will be on the die also. Not that I think it's an issue unless you bolt mod the console or increase spring tension (not recommended).

 

[D777]

Errorcodes from ru forum

Spoiler

A0802131 CPU/GPU overheating doesn't actually mean anything.
A0203010 southbridge/hub failure or failure
A0403034 this error means poor contact or no contact at all, you need to reball gpu, south or processor.
A0805FFF presumably dead processor
A0003001 power supply error, possibly a short circuit, the CPU may also be faulty. I received it on a board with a soldered processor.
Errors after incorrect CPU scalping:
OK 00000000 A0404421 FFFFFFFF:4A
OK 00000000 A0403034 10D36ED5:E5
OK 00000000 A0801601 0B733199:CD
OK 00000000 A0801701 0B733199:CE
OK 00000000 A0802131 0B6FE9D6:00
OK 00000000 A0802130 0B6F9D66:F0
and one more option:
OK 00000000 A0404402 0B4AC585:E5
OK 00000000 A0801601 0B4AC56E:F5
OK 00000000 A0801701 0B4AC56E:F6
OK 00000000 A0403034 0B49D953:
OK 00000000 A0801802 0B49D914:F

The CPU / GPU blade or the balls stuck together under them may also be faulty:

OK 00000000 A0402030 0B49E9A2:E4
OK 00000000 A0402131 0B49E99D:F0
OK 00000000 A0402031 0B49E99C:EE
OK 00000000 A0404441 FFFFFFFF:4C
OK 00000000 A0403034 FFFFFFFF:49
OK 00000000 A0003001 FFFFFFFF:3F

GPU or CPU dump, also possible gray nickels under them, reball is required
OK 00000000 A0801802 0B49D914:F
OK 00000000 A0801701 0B4AC56E:F6


Errors with a soldered south bridge may also indicate its malfunction:
OK 00000000 A0093003 FFFFFFFF:4A
OK 00000000 A0091005 FFFFFFFF:4A
OK 00000000 A0081005 FFFFFFFF:49
and one more option:
OK 00000000 A0404441 FFFFFFFF:4C
OK 00000000 A0403034 FFFFFFFF:49


If you forgot to connect the power supply or a short circuit on the board (for example, the balls stuck together), the error codes may look like this:
01:6C
03:6E
04:6F
02:6D
Brick PS3, operating system problem, firmware crashed
All hands do not reach to deal with this console, maybe I will create a separate topic on the forum on it, I threw it during the firmware via HAN or HEN.
OK 00000000 A0404432 FFFFFFFF:4C here is the main error indicating a problem with the firmware, but until I restored this prefix I will not mark it as found.
OK 00000000 A0403034 FFFFFFFF:49
OK 00000000 A0801601 FFFFFFFF:4B
OK 00000000 A0801701 FFFFFFFF:4C
OK 00000000 A0802022 1A4595BF:EA

 

[Thibobo]

Pay attention to the preheater working area. The AHCI SC Pro has a much bigger area than similar looking ones.

Thanks, i'm also wondering if the rsx is delidded, should we reseal it by applying thermal glue to the ram modules on the rsx and placing back the ihs? How did sony do this in their "official" frankenstein consoles?

 

[RIP-Felix]

Thanks, i'm also wondering if the rsx is delidded, should we reseal it by applying thermal glue to the ram modules on the rsx and placing back the ihs? How did sony do this in their "official" frankenstein consoles?

Yes, I reccomend stars922 thermal plaster (on RSX). Sony didn't replace the paste. They used new old stock RSX which are designed to be reflowed during the initial install. When they got overheating CPUs they cheated and used a CELL bracer. They didn't delid either.

 

[Thibobo]

Yes, I reccomend stars922 thermal plaster (on RSX). Sony didn't replace the paste. They used new old stock RSX which are designed to be reflowed during the initial install. When they got overheating CPUs they cheated and used a CELL bracer. They didn't delid either.

Thanks once again for your info!

What exactly is a "CELL bracer"? , i'm also wondering where the best place is to source the BD3504 voltage regulator other than harvesting it from a slim ps3?
Or maybe there is another chip which does the same thing also msop8 size?
I found this but seems very expensive:

 

[RIP-Felix]

I got new old stock BD3504s on ebay. Each COK-001 has 2 and COK-002 has one. Slims dont have one. No alternative I'm aware of.

Cell bracer:

 

[Thibobo]

I got new old stock BD3504s on ebay. Each COK-001 has 2 and COK-002 has one. Slims dont have one. No alternative I'm aware of.

Cell bracer: View attachment 40740

Cool thanks! What did they want to achieve with that piece of uhm plastic..

Is there any 3D printer file available to print that piece ourselves?

I'll be looking out for those BD3504 ic's or search for a alternative that would fit the pinout on the motherboard.
I found this datasheet that might be useful for finding an alternative: https://pdf1.alldatasheet.com/datasheet-pdf/view/336717/ROHM/BD3504FVM.html


Another question i have, the "orbis mod chip" is there any documentation about that chip, like the pcb gerber & ic used? I want to make one myself, but can't find this info anywhere...

Thanks once again for your time

 

[RIP-Felix]

I made a STL, but haven't released it. Lots of other side projects.

It braces the bottom of MB under the Die, preventing the MB under the BGA from bowing, or "tacoing." This restores contact with the TIM without delidding. But it's not a long term solution. The paste will continue to deteriorate and delidding will become necessary. So at best, sony put a band aid on it.

Please note that the bracer DOES NOT ADD PRESSURE. It is exactly 1mm thick, the same thickness as the 4 standoffs on the RF shield backplate that the leaf springs screw into. All it does is "brace" more of the motherboard underneath the BGA so that the board cannot sag. No upward force is applied. It keeps the MB flat (coplaner). Serving 2 jobs.

1. maintaining uniform bondline thickness with the paste for optimal thermal contact (this is a fact).
   
2. stiffening of the BGA package to reduce warping stresses that could weaken the joint (this is my opinion).

ORBIS was created by Kiaw (@botakompong's brother). He passed away and AFAIK only botakompong still has the original design files. He sells the modchip, so I guess it's closed source. Perhaps he will consider making it open source when he retires. But that's up to him.

 

[Thibobo]

I made a STL, but haven't released it. Lots of other side projects.

It braces the bottom of MB under the Die, preventing the MB under the BGA from bowing, or "tacoing." This restores contact with the TIM without delidding. But it's not a long term solution. The paste will continue to deteriorate and delidding will become necessary. So at best, sony put a band aid on it.

Please note that the bracer DOES NOT ADD PRESSURE. It is exactly 1mm thick, the same thickness as the 4 standoffs on the RF shield backplate that the leaf springs screw into. All it does is "brace" more of the motherboard underneath the BGA so that the board cannot sag. No upward force is applied. It keeps the MB flat (coplaner). Serving 2 jobs.

1. maintaining uniform bondline thickness with the paste for optimal thermal contact (this is a fact).
   
2. stiffening of the BGA package to reduce warping stresses that could weaken the joint (this is my opinion).

ORBIS was created by Kiaw (@botakompong's brother). He passed away and AFAIK only botakompong still has the original design files. He sells the modchip, so I guess it's closed source. Perhaps he will consider making it open source when he retires. But that's up to him.

Thanks for the info once again!

Another question that sparked my mind is what thermal silicone you'd recommend to reseal the CELL ?

 

[RIP-Felix]

That's an issue. I have not found one I like that equals the properties of the OG. I have sucessfully used RTD silicone gasket maker. @squeept has been using it sucessfully for many years. But it's softer. Much softer. While it retains the IHS, I doubt it provides the stiffness to protect the BGA in bullet point #2 above. I have not been able to find an appropriate IHS silicone material. Tho I've not gone to great effort to find one either. I'm sure it exists, it's just there's so much more information about delidding than there is about regluing, that it buries the search results under hundreds of useless pages only telling how to delid. They always stop there and almost never talk about re-adhering it afterwards. And when they do, use the wrong stuff. So yeah, it's an issue.

 

[santorfo]

Hey guys, I have a PS3 Slim CECH-2004B that over the course of I'd say 2 months started showing the usual symptoms, sudden shutoffs after the console was on for a bit, that I thought were overheating at first, then I thought I'd found a way to mitigate them by first playing a less demanding game for a bit to "warm up" the console then jumped on a more demanding game and it worked for a while then eventually it started happening more and more, the "warm up" wasn't working and eventually I got into the red blinking light state every time I tried powering on the PS3, after about 5 seconds.

Spoiler: Potentially useless backstory

This was all in 2019 I believe. After the immediate shutoffs started happening I left the console alone for a long time thinking that it was dead. When RIPFelix posted his YLOD history video in September 2022, it got to my recommended pretty early on (it had something like 4k views) and I was obsessed for a while with the issue, especially the bumpgate issue cut off which made me almost certain that my 65nm RSX kicked the can because the manfacturing dates lined up.
After this, I went back to the attic and tried turning the PS3 on. Sure enough, the same symptoms ocurred but I kept trying. I sat it on its side, I held it upside down, and then eventually sat it on its topside, in a frustrated attempt to get a boot. To my absolute shock it actually booted into the XMB and I left it alone for like 10 minutes to see if it would stick. Came back, it was still on. I used webman to check temps and a bunch of other stuff and it seemed like the console was on fine. I didn't play any games and I was sure that as soon as I turned it off it wouldn't turn back on again, and that's exactly what happened. Defeated, I put it back in the attic and it remained there for a few more months.

So when I found out about the SYSCON method, I got into soldering. Fixed some other stuff, like cables and controllers first and then opened my PS3 up for the second time ever. I soldered the RX and TX leads and sure enough I got my SYSCON codes:

Spoiler: SYSCON codes

Microsoft Windows [Version 10.0.22000.1696]
(c) Microsoft Corporation. All rights reserved.

C:\Users\roliveira\Downloads\PS3-main\Syscon Connection Drivers and Diagrams ByNSC>python "PS3_Syscon_(Mullion_Sherwood)_script.py" COM4 SW
> AUTH
Auth successful
> ERRLOG
00000000
# CODE CLOCK
# A0801002 15A8FF8E
# A0801002 15A8FF74
# A0801002 15A8FF5F
# A0801002 15A8FF3A
# A0801002 15A8FF29
# A0801002 15A8FF16
# A0801002 15A8FF01
# A0801002 15A8FEEC
# A0801002 14082CF6
# A0801002 14082CD2
# A0801002 14082CC2
# A0801002 14082CB5
# A0801002 14082CA1
# A0801002 14082C6E
# A0801002 14082C5D
# A0801002 14082C47
# A0801002 14082C2A
# A0801002 134A616D
# A0802022 134A613B
# A0802022 134A6139
# A0801002 134A5B9B
# A0801002 134A5AED
# A0801002 134A5665
# A0801002 134A5639
# A0801002 134A5620
# A0801002 134A53BD
# A0801002 134A530E
# A0801002 134A529D
# A0801002 134A5260
# A0801002 133D3BC4
# A0801002 133D3BB2

> ERRLOG GET 00
00000000 A0801002 15A8FF8E
> ERRLOG GET 01
00000000 A0801002 15A8FF74
> ERRLOG GET 02
00000000 A0801002 15A8FF5F
> ERRLOG GET 03
00000000 A0801002 15A8FF3A
> ERRLOG GET 04
00000000 A0801002 15A8FF29
> ERRLOG GET 05
00000000 A0801002 15A8FF16
> ERRLOG GET 06
00000000 A0801002 15A8FF01
> ERRLOG GET 07
00000000 A0801002 15A8FEEC
> ERRLOG GET 08
00000000 A0801002 14082CF6
> ERRLOG GET 09
00000000 A0801002 14082CD2
> ERRLOG GET 0A
00000000 A0801002 14082CC2
> ERRLOG GET 0B
00000000 A0801002 14082CB5
> ERRLOG GET 0C
00000000 A0801002 14082CA1
> ERRLOG GET 0D
00000000 A0801002 14082C6E
> ERRLOG GET 0E
00000000 A0801002 14082C5D
> ERRLOG GET 0F
00000000 A0801002 14082C47
> ERRLOG GET 10
00000000 A0801002 14082C2A
> ERRLOG GET 11
00000000 A0801002 134A616D
> ERRLOG GET 12
00000000 A0802022 134A613B
> ERRLOG GET 13
00000000 A0802022 134A6139
> ERRLOG GET 14
00000000 A0801002 134A5B9B
> ERRLOG GET 15
00000000 A0801002 134A5AED
> ERRLOG GET 16
00000000 A0801002 134A5665
> ERRLOG GET 17
00000000 A0801002 134A5639
> ERRLOG GET 18
00000000 A0801002 134A5620
> ERRLOG GET 19
00000000 A0801002 134A53BD
> ERRLOG GET 1A
00000000 A0801002 134A530E
> ERRLOG GET 1B
00000000 A0801002 134A529D
> ERRLOG GET 1C
00000000 A0801002 134A5260
> ERRLOG GET 1D
00000000 A0801002 133D3BC4
> ERRLOG GET 1E
00000000 A0801002 133D3BB2
> exit

C:\Users\roliveira\Downloads\PS3-main\Syscon Connection Drivers and Diagrams ByNSC>python "PS3_Syscon_(Mullion_Sherwood)_script.py" COM4 SW
> auth
Auth successful
> errlog
00000000
# CODE CLOCK
# A0801002 15A8FF8E
# A0801002 15A8FF74
# A0801002 15A8FF5F
# A0801002 15A8FF3A
# A0801002 15A8FF29
# A0801002 15A8FF16
# A0801002 15A8FF01
# A0801002 15A8FEEC
# A0801002 14082CF6
# A0801002 14082CD2
# A0801002 14082CC2
# A0801002 14082CB5
# A0801002 14082CA1
# A0801002 14082C6E
# A0801002 14082C5D
# A0801002 14082C47
# A0801002 14082C2A
# A0801002 134A616D
# A0802022 134A613B
# A0802022 134A6139
# A0801002 134A5B9B
# A0801002 134A5AED
# A0801002 134A5665
# A0801002 134A5639
# A0801002 134A5620
# A0801002 134A53BD
# A0801002 134A530E
# A0801002 134A529D
# A0801002 134A5260
# A0801002 133D3BC4
# A0801002 133D3BB2

So it seems to be A0801002, which, from what I can tell from my short reading of this thread, is usually the fault of the tokins. I knew that replacing the tokins had a chance of bringing it back to life but I never bothered to actually try it because I thought I should take more precautions and actually ask in here first before doing it. So here I am. Is it worth it? Slim PS3s where I live sell for around 50€ and I wouldn't mind buying one but I would really like to give it a shot and try to repair it, it's not about the money, it's about the pride of (maybe) being able to fix it, even if it's only for a short while.

I found post #3546 in page 178 from @Pacorretaco in which he soldered what I think is a tantalum capacitor? to the tokins and didn't replace any of the original ones and got a CECH-20XX to work again, so if that's something that I could maybe do then that seems like the least destructive way to attempt to fix this console. Thanks in advance and thank you Felix for all the investigation and help you've given to people regarding these issues, the world needs more people like you.